JDBC driver does not support compressed IPC streams (ZSTD, LZ4) [FlightSQL]

### Describe the bug, including details regarding any error messages, version, and platform.

**Component:** Flight SQL JDBC Driver
**Version:** 18.3.0

The `flight-sql-jdbc-driver` shaded JAR does not include compression support, causing connection failures when the Flight SQL server sends compressed IPC streams using ZSTD or LZ4 compression.

### Expected Behavior
The JDBC driver should be able to handle compressed IPC streams from Flight SQL servers, similar to how the direct Flight SQL client (`flight-sql` artifact) handles them.

### Actual Behavior
Connection fails with one of these errors depending on compression type used by server:

**ZSTD compression:**
```
java.lang.IllegalArgumentException: Please add arrow-compression module to use CommonsCompressionFactory for ZSTD
at org.apache.arrow.driver.jdbc.shaded.org.apache.arrow.vector.compression.NoCompressionCodec$Factory.createCodec(NoCompressionCodec.java:63)
```

**LZ4 compression:**
```
Arrow error: Invalid argument error: lz4 IPC compression requires the lz4 feature
```

### Steps to Reproduce

1. Set up a Flight SQL server that uses compression (ZSTD or LZ4)
2. Try to connect using the JDBC driver:
String url = "jdbc:arrow-flight-sql://localhost:5127?useEncryption=false";
Connection conn = DriverManager.getConnection(url);
Statement stmt = conn.createStatement();
ResultSet rs = stmt.executeQuery("SELECT * FROM table");
// Fails with compression error
### Workaround
Currently, either:
1. Disable compression on the server side, OR
2. Use the direct Flight SQL client API (`flight-sql` and `flight-core` artifacts) instead of JDBC

### Additional Context
- Direct Flight SQL client works fine with compression
- This affects any Flight SQL server that enables compression by default
- JDBC drivers should be able to handle all protocol features transparently

首先比较 flight-sql-jdbc-driver 的 shaded JAR 中包含的依赖项与 flight-sql 和 flight-core 工件，然后使用 ZSTD 或 LZ4 压缩重现与服务器的 JDBC 连接。完成标准是 JDBC 驱动程序能够连接并读取结果，且不会出现压缩错误。

由索引模型根据 Issue 内容生成。